Well, obviously, my posts are largely speculative when it comes to something that doesn't exist (for modern drives) or that existed - in a private and publicly undocumented form - for drive technology in the past.
I am happy to hear that your apparatus read 100 sectors per second at the time, let's say that todays processors are 40 times faster than the whatever you used, but you'll have to concede that an hypothetical today's apparatus - and provided that the *may* is actually a *can* - designed to read last generation drive's platter's will have to read much more tiny data (or head misplacement or deviation from "theoretical alignement" or whatever is the proper name).
If I am not mistaken the "vertical recording" was introduced to overcome a physical limit, today hard disks' heads are (or are advertised as) much more "exact" in placement that before, in order to be able to read properly much more compactly recorded platters.
Thus I presume that the x40 factor above should be reduced.
So maybe we could use on my "60 sectors per second" base
60*40=2400
100*24=2400
However the probabilities do not fall that much, they go from a
<very high number, in the millions range>12
or
<very high number, in the millions range>1
to
<very high number, in the millions range>480
or
<very high number, in the millions range>40
Unless the factor I used is grossly incorrect, of a couple orders of magnitude, the chances are really low.
On the other hand, there is a psychological factor that might be introduced
- criminals tend to be really good at playing with odds
- caring, conservative people tend to be not criminals
All in all I would say
- if you are a criminal, you may want to take the odds and do a single wipe
- if you are not a criminal, you have no reason in the world to do anything more than a single pass
- if you are paranoid, or dumb, or both, you should do all the 35 passes voodoo, then apply a strong magnetic field to the platters and a voltage burst to the PCB, and finally destroy the drive in a hydraulic press, before burning it with napalm and blowing it up with not less than 350 grams of C4 wink , even then, if the process was filmed with a high speed camera and recorded with a very high fidelity audio equipment, CIA or NSI could reproduce the previous contents of the hard disks 😯
In other words, in 99,9999% of cases, a single pass is enough for practical uses, as I see it.
And as always, I would love my approach, or the conclusions that come from it to be discussed ), and if it's the actual case, be proved wrong.
jaclaz
Jaclaz
You seem intent on ignoring the massive parallelism of the above task - one the platters are read the task is simply a computing one and one that is carried out with much faster processors. Add multiple processors and you could easily be doing many thousands of sectors per second - the throughput would simply be limited by your budget.
That being said I agree that a single pass enough for every day use - but I said that in my first post.
I am giving up on tihs now though - its starting to go around in circles )
Jaclaz
I am giving up on tihs now though - its starting to go around in circles )
I have been following this quite intently and agree 99.9% single pass is more than sufficient so I say there's no point flogging a dead horse any more … let alone flogging a dead hard drive 😉
Do we actually have any figures on the likelihood of misregistration to any useful degree on modern hard drives?